This study used a single-subject design to determine the effectiveness of the Cool versus Not Cool (CNC) discrimination procedure to teach social skills to three young U.S. adults with autism. The CNC procedure was administered remotely during this study. The specific social skills targeted were asking open-ended questions and responding appropriately to a conversation partner's answers. All participants significantly increased the accuracy of the targeted social skills after receiving the intervention. The skills also maintained at relatively high levels once the intervention was removed. Future research and implications are discussed.